Amazon told the public that the 4-day Prime event surpassed any 4-day period in its history that included a shopping event. I’d hope so. This is the first time they’ve extended this event from 2 to 4 days, so not surpassing previous records would have been significantly concerning. Good to hear that didn’t happen. Perhaps more importantly, Adobe reported 30.3% e-commerce industry growth over that period of time, which is materially better than its 28.4% estimate. Considering Amazon owns nearly half of the U.S. ecommerce market and consistently takes market share, it’s reasonable to assume outperformance for everyone likely means solid results for Amazon as well. EPS estimates for 2025 have ever-so-slightly ticked higher since the event, but candidly, we could call that positive movement a rounding error.

CrowdStrike added its Falcon Model Context Protocol (MCP) and CrowdStrike AI Red Team (managed security help) to its AI Agents and tools offering on the AWS marketplace. AWS was the first channel partner to deliver more than $1B in cumulative sales to CrowdStrike, and this should simply bolster that momentum.

Ben Thompson (@benthompson) wrote a great article this week and I wanted to paraphrase the key point. Cloudflare’s AI crawler blocking practices help companies safeguard data from large language models (LLM) and agent data scraping. Alphabet uses its own crawler (Googlebot) for AI Mode and AI Overviews. So? If a merchant wants distribution on Alphabet’s properties, they must grant access to Googlebot. This means most agents and models don’t have access to about a fifth of the world’s internet traffic – but Alphabet does. This should deepen its data advantage vs. the field. Apps/models are only as good as the data they can use. And considering it already has distribution to 2B people for 6 products, this will merely extend its already large lead.

Demand:

Bank of America missed revenue estimates by 1%.

Margins, Returns & Capital Ratios:

The company beat $0.86 EPS estimates by $0.03. Outperformance in lending and asset trading drove the beat.

Needed Credit Definitions:

Commentary on the Credit Data Below:

Outlook & Valuation:

Bank of America continues to expect quarterly net interest income to ramp to $15.6B by the end of the year and about 6.5% Y/Y net interest income growth.

The company trades for 12x forward EPS. EPS is expected to grow by 14% this year and by 16% the following year.

Important Call Commentary:

“We continue to see a solid consumer spending data, improving credit quality from already strong statistics and plenty of household net worth growth…  that leads our worldwide leading research team to continue to predict no recession, a modestly growing economy about 1.5% at the end of the year and continued no Fed rate cuts till next year.”

CEO Brian Moynihan

“Focusing on total net charge-offs again and looking forward, in the near term, we would not expect much change in the total net charge-off ratio given the steadiness of consumer delinquencies, stability of C&I and the reductions in our CRE office exposures.”

CFO Alastair Borthwick

Demand:

Revenue was 2.0% ahead of expectations. Net interest income growth excluding markets fell by 1% Y/Y via rate cuts.

The overall Y/Y revenue decline was solely driven by the corporate revenue segment falling from $10.1B to $1.54B Y/Y This was due to a large boost from their Visa equity stake and a $2.7B First Republic bargain purchase gain. All core operating segments grew Y/Y. Excluding the corporate segment, revenue grew by 8.2% Y/Y.

Margins, Returns & Capital Ratios:

Commentary on the credit data below:

Everything looks reasonably good, which is great news for the U.S. economy’s overall health.

Outlook & Valuation:

The company raised net interest income (NII) guidance for the year by 1.1%. Excluding the volatile markets segment, it raised annual NII guidance by 2.2%. Finally, it maintained 3.6% card service net charge-off expectations for 2025.

JPM trades for 15x forward EPS. After 2 years of EPS compounding at a 28% clip (related to those Visa and First Republic Gains), EPS is expected to fall by 2% Y/Y this year and grow by 6% the following year.

Important Call Commentary:

“The U.S. economy remained resilient in the quarter. Tax reform and deregulation are positives for the outlook. Still, significant risks, including from tariffs, geopolitics, deficits and elevated asset prices, persist.”

CEO Jamie Dimon

“When it comes to tariffs, there's now more talk as more things get done. Some tariffs have been delayed, which reduces the risk a little bit. So there's still risk out there, but I am hopeful that some of these frameworks will be completed before August 1.”

CEO Jamie Dimon

“We look at [consumer health] very closely. It obviously matters a lot for us as a company. But we continue to struggle to see signs of weakness… you see a bit more stress in lower income bands, but that’s always true.”

CFO Jeremy Barnum

“Delinquency rates are in line with expectations. That all looks fine… while there are nuances around the edges, consumer credit is about labor markets. In a 4.1% unemployment rate world, it’s going to be hard to see a lot of weakness.”

CFO Jeremy Barnum

“If you look at government data, first half consumer spending vs. second half of last year is down. It’s still growing, just more slowly. That’s consistent with the soft landing narrative. Our own data on a nominal basis actually shows spending a bit up over that period.”

CFO Jeremy Barnum

Demand:

American Express beat revenue estimates by 0.8%.

Profits, Returns & Margins:

Guidance & Valuation:

American Express reiterated annual guidance calling for 9% revenue growth, which beat 8.1% Y/Y growth estimates. It also reiterated $15.25 in EPS (14% Y/Y growth), which met estimates.

AXP trades for 18x forward EPS. EPS is expected to grow by 14.2% Y/Y this year and by 13.7% next year.

Credit Data & Commentary:

As you’d expect for an ultra-prime lender like AXP, their credit data looks very good.

Call Commentary:

“When you look at our performance across spend, transactions, retention and credit in the context of the significant macroeconomic and geopolitical developments of the past few months, what you see is remarkable resilience across our customer base.”

CFO Christophe Le Caillec 

“Delinquency and write-off rates remained low... Our strong credit performance is remarkable across all age groups.”

CFO Christophe Le Caillec 

“U.S. Consumer millennial spend was up 10% and GenZ grew by 40%. International continued to grow in the double-digits.”

CFO Christophe Le Caillec

“Total Card Member spending was up 7%, which was consistent with the pattern we've seen this year, while spend in some of the travel categories like airlines and lodging was softer overall, spending was a quarterly record.”

CEO Stephen Squeri

“I think from an SMB perspective, while billings are probably not where we want them to be, our revenue from this segment continues to be strong. Our credit metrics continue to be strong. Our lending book is strong and our fee-based business here is strong. So overall, we're happy with SMB.”

CEO Stephen Squeri

“Credit card spending growth softened very slightly in the second quarter, but is still up year-over-year and remains strong overall, and debit card spending growth has remained strong and consistent with what we saw in prior quarters.”

Wells Fargo CEO Charlie Scharf

“Consumer delinquencies continued to improve from a year ago and commercial credit performance continued to be relatively strong… Consumers and businesses remain strong as unemployment remains low and inflation remains in check.”

Wells Fargo CEO Charlie Scharf

“The strength of the U.S. economy, driven by the American entrepreneur and a healthy consumer has certainly been exceeding expectations of late”

Citi CEO Jane Fraser

The clear theme of this week’s bank and credit earnings was economic resilience. Whether it’s consumers, small businesses or large enterprises, the U.S. economy continues to be healthy and support needed spending growth. Credit metrics also look good. Considering these institutions have arguably the best view of the macro backdrop, hearing uniformly positive things about it is certainly encouraging.

We need to update this chart from Uber’s Q4 earnings. Since it was shared a few months ago, May Mobility, Momenta and Mobileye (through a Volkswagen arrangement) have all joined the partner roster. And this week, Baidu was added to the list, as the remaining Commercial L4 autonomy holdout joins forces with Uber. This year, Baidu will deploy its Apollo Go AVs through Uber’s app in cities across the Middle East and Asia. Over the coming years, this partnership will scale to thousands of vehicles, with Uber’s unmatched demand aggregation talents ensuring the expensive hardware is optimally utilized and profitable. Apollo Go has already crossed 10M rides globally, and now as it continues to pursue global scale, it will do so hand-in-hand with Uber.

But wait, there’s more. Uber is partnering with Lucid to deploy 20,000 or more AVs over the next 6 years. The first launch will be in a “major U.S. city” in 2026. Lucid is a cash incinerator without much product-market fit as of today – but it does have a car with a compelling 450 mile range that could be ideal for AV programs (less frequent charging). Considering this, I was very glad to hear that Nuro will provide the autonomy software needed to power these cars. Nuro has been an Uber partner for a while, and has achieved level 4 autonomy. The first Nuro/Lucid model is already being tested in controlled environments. These cars will be owned by Uber or its fleet managers and only available on Uber’s app. As part of this news, Uber is investing $300M in Lucid and reportedly a bit more than that in Nuro.

Moreover, I love the Baidu partnership. It shows the leaders in this space continuing to flock to Uber’s platform and also provides more evidence of there being many, many future players in this space. As I’ve argued for a long time, market fragmentation is how Uber’s network effect and demand aggregation will be most valuable. If one company owns the entire market, that network effect becomes irrelevant and the monopoly can easily offer a product with zero substitutes to attract demand. Based on this news, as well as several deployments happening from companies not named Waymo or Tesla through 2026 and Tesla’s slow, controlled Austin rollout… the market fragmentation prerequisite continues to become more likely. 

The Lucid deal is more interesting. Uber has long talked about owning some cars early in this transition to alleviate supply bottlenecks and ensure it’s part of the evolution. I’m not a fan of  securing rights to future cars from a pre-profit company that has struggled since going public, but the cash infusion will be earmarked towards manufacturing those specific units and supporting Nuro integration. And I am excited about the Nuro investment. They’re a highly-regarded software vendor in this space, and clearly are able to integrate their tech into non-AV models like the Lucid Gravity. That means more partnerships with auto manufacturers can be struck and more of these manufacturers can offer autonomy-enabled models without needing to develop that capability internally. This bodes well for more market fragmentation, as the barrier to entry falls with news like this and Nvidia’s help with synthetic data generation to greatly lower the cost of training these models.

As we’ve spoken about recently, Argentina has flipped from the geographic black eye in MELI’s performance to the market that’s carrying its results. Hyperinflation there has begun to cool meaningfully, economic growth is ticking higher and credit ratings are rising. And the revenue uptick stemming from this should also be margin accretive. This is MELI’s highest-margin market, and macro there is finally starting to look a lot more robust. Perhaps in response to this, it’s investing another $65M in storage capacity. This follows construction of its second fulfillment center in that country currently being built.

Under new Chief AI Officer Alexandr Wang, Meta is considering a shift to closed-source models. This could mark a sharp pivot from the open-source approach it has been so loud about and confident in. Why is this happening now? Meta has always thought of open sourcing as a way to make sure world-class developers are building on its platform. World-class distribution and great models are a wildly tempting combination to attract developers. 

But now? Over the last several weeks, Meta has been spending like crazy to steal top developer talent away from OpenAI (two more hired this week), Anthropic, Apple, Alphabet and everyone else. It’s buying all of the talent it needs to make sure needed and effective work is done on Llama, without needing to rely on these builders choosing its product. It no longer feels the need to open source Llama to ensure it stays competitive, as now, it’s more confident in building proprietary models internally.

Zuck is also passionately determined to ensure this elite talent has all of the compute capacity they could possibly need. They’re building “several” compute clusters with multiple gigawatts in scale, with one of these apparently planned to scale to 5 gigawatts. 

As a quick and relevant aside, when Apple restricted cross-app data sharing, Meta’s business suffered. Zuck made a commitment to plugging the signal gap Apple created with AI and shedding reliance on future platforms for distribution. How do you do that? Having the best AI talent… with access to massive amounts of compute… while boasting a convincing lead in AI hardware and enjoying ubiquitous distribution. They have everything they need to ensure they’re fully in charge of their own destiny.

DraftKings has pulled its prediction market application and is reportedly exploring a purchase of Railbird Exchange. This is a smaller player in the budding space, and would likely be integrated into the overall DraftKings brand and app to bolster bet breadth. As a reminder, I view prediction market legality as a large net positive for DKNG. If ESPN, Barstool, casinos and others couldn’t make a dent in DraftKings/Fanduel domination, I don’t think Kalshi or others will either. But even if they take a little volume, DraftKings unlocking this type of betting would naturally foster incremental revenue for it too. Furthermore, regulatory clarity would mean DKNG can potentially enter the rest of the states without legal betting and could enjoy significant tax relief in existing states too. They’d also have more flexibility to direct excess bet volume for certain lines to this format to cater to more demand without more risk.

The student loan payment pause is ending. Students will now have a short period of time to get current on their payments as interest accrues. This should greatly support refinancing application volume. SoFi owns more than 60% of this market for its own credit bands and should enjoy a lot of incremental demand here. Great timing, considering more near future rate cuts should support refinancing interest even more. SoFi has delivered steady top-line growth for years while this business has remained largely dormant. That’s despite student loans being SoFi’s largest revenue driver when the moratorium initially began. It’s exciting to think what this already healthy financial engine will look like with yet another tailwind to enjoy. 

One other note on this. Some argue that student loans are not compelling revenue compared to personal loans and shouldn’t be pursued. I don’t agree. While it’s true that net interest margin (NIM) is higher for personal loans, so are net charge-offs. Personal loan borrower quality is lower compared to student loans and this is a great way to grow the balance sheet and collect more net interest income without taking quite as much risk. 

But they may not even need to make this balance sheet growth and risk trade-off. SoFi has openly talked about opening up its student loan business to lending platform partners. That would mean student loan originations don’t drive balance sheet growth and don’t create any kind of bottleneck for higher NIM personal loan capacity. Personally, I’d love to see that happen, but still think student loan revenue should be pursued even if it entails using the balance sheet. Student loan profitability is still far more compelling than holding cash and equivalents. So swapping those assets for student loans is still margin accretive and SoFi still has plenty of capital ratio cushion to utilize. Aside from all of this, student loans are a great top-of-funnel product to support member growth and more cross-selling.

Piper Sandler digital advertising channel checks are in. For Meta, while cost per 1,000 impression growth slowed, cost per click and engagement were notable strengths. Sounds like targeting is getting better and placements are getting more valuable as impressions more meaningfully translate into interested buyers and more profitable campaigns. They observed strong social media ad market share and think Asia-Pacific headwinds like Temu cutting ad spend to zero are priced in. They’re upbeat about the quarter. For Alphabet, they observed some modest declines in search advertising market share, but aren’t overly concerned as they take a neutral stance on earnings. Some advertisers saw issues with Alphabet advertising during Q1, but “they’ve since regained momentum.”

Bank of America Q2 digital advertising channel checks were also positive. Following the Liberation Day-inspired April slowdown, demand recovered and normalized. Alphabet enjoyed steady search growth, with that expected to carry into Q3. They expect Alphabet to modestly surpass profit estimates, while channel checks for the cloud business were also quite positive. Meta enjoyed a modest Q2 acceleration that prompted small raises to BofA’s estimates for the company. For Amazon, the analyst cited a modest Q2 slowdown, while Snapchat materially slowed during the period. Pinterest accelerated, but carries more tariff-related risk than the other social platforms.

Keybanc Capital Markets channel checks revealed strong digital ad spend for Meta during Q2. They expect a beat and raised their estimates for both 2025 and 2026.

Morgan Stanley downgraded CrowdStrike to equal-weight. This has nothing to do with how the business is performing. Fundamental momentum is excellent as the company continues to admirably overcome last year’s outage. Expectations are just sky-high and the valuation is too. This is similar to some of the other notes we’ve seen in recent weeks, and is something I find very fair. I think a breather would be healthy for this name.

Cowen initiated SoFi with a hold rating. They are attracted to SoFi’s prime credit niche, cross-selling engine, and loan platform business growth. They’re also (rightfully) not overly concerned about its usage of fair value accounting (as they shouldn’t be). Still, after the explosive recent run, they view shares as fairly valued for now.

Cowen reiterated an outperform rating for Uber. They see bookings actually coming in below expectations but profit coming in ahead.

Cowen sees a strong quarter coming for Microsoft and Azure specifically. They think Azure will keep accelerating and other products will modestly outperform expectations.

Morgan Stanley sees a rock-solid quarter for Cloudflare coming based on survey data. They continue to effectively innovate and cross-sell more products. More activity with channel partners was also cited as a positive, as they’re enjoying a faster cadence of large deal wins.

Wedbush published significant survey data from cybersecurity channel partners. CrowdStrike performance was about 7% better than expected. Palo Alto & Zscaler were both around 3.5% ahead of expectations. SentinelOne continues to struggle. Win rates on big deals fell vs. CrowdStrike, but did rise in mid-market and smaller deals vs. other competitors. All in all, the company’s channel partner performance was about 3.5% worse than expected.

Wedbush sees Nvidia H20 sales raising 2026 EPS by roughly $0.38 or 9.2%.

New deals for Oracle prompted Cantor to raise their estimates for 2026 and 2027. Oracle Cloud Infrastructure and its suite of database products are all thriving.

Per Wedbush, checks point to a materially outperforming quarter for CrowdStrike. Win rates vs. SentinelOne and Microsoft Defender rose. Palo Alto and Zscaler channel checks were modestly positive, but not as upbeat as for CrowdStrike. It looks like SentinelOne will again deliver an underwhelming quarter. 

Wedbush sees an “encouraging near-term setup” for The Trade Desk. They’re more confident in TTD’s competitive positioning and the noisy Kokai launch. Their surveys also lead them to think there could be Q2 upside vs. consensus expectations. Meta & Alphabet survey data and channel checks were similarly positive. For Alphabet, they see compelling upside potential for rest-of-year results. Advertising demand sharply recovered following a blip of weakness during the trade drama in April.

Deutsche Bank channel checks are pointing to a Q2 beat for Shopify. They expect outperforming volume, revenue and profit, with channel checks pointing to market share gains and Liberation Day resilience.

Like everyone else, Keybanc digital advertising data was positive for Q2. They expect strong results for Alphabet, Pinterest and Roku. They also expect good data from Meta and The Trade Desk (welcome to the S&P 500), but think somewhat polarizing bull/bear debates won’t be resolved from these quarters. They think that The Trade Desk is “back on track” in terms of execution, which meshes well with other recent notes citing improving Kokai performance and user satisfaction. They raised estimate data for Meta, Alphabet, Pinterest (aside from lowering MAUs) and The Trade Desk.

Seaport upgraded PayPal from sell to neutral. This was based on brighter-than-expected tariff outcomes, improving growth and operational execution driving higher estimates.

Truist maintained a hold rating on SoFi. They expect a great quarter, but think the explosive run up in shares has elevated expectations too much. They said basically the same thing about Affirm too.

UBS sees accelerating growth for Chipotle in Q3 and Q4. Comps are getting easier, macro is better, menu releases are doing well, marketing returns are strong and traffic is rising. They continue to have confidence in 20% compounded EPS growth over the coming years.

Jefferies downgraded Starbucks to underperform. It thinks the stock is too expensive and the turnaround will take more time than expected.

Bank of America downgraded SentinelOne due to execution issues. I think this is fair. They have a ton to prove and need to start delivering better results. On the other hand, Rosenblatt initiated coverage of SentinelOne with a buy rating this week. They think AI product traction is building, non-endpoint cross-selling is gaining momentum and the stock is undervalued vs. peers. It will be a very interesting quarter for this company.

Great macro data week. June is when tariffs were supposed to start showing up in inflation data, and the readings we got were quite positive. Employment metrics were strong. Output metrics were strong. Everything was strong.

Inflation Data:

Output Data:

Consumer & Employment Data: